A Lightweight Convolutional Neural Network-Based Method for Cotton Mosaic Disease Identification
Bibliographic record
Abstract
Common cotton diseases like cotton leaf bright, leaf curl, and fusarium wilt have a direct impact on the final production and quality of cotton. the reduction of economic losses, and the promotion of the cotton industry may all be achieved via prompt disease prevention and accurate diagnosis. Correct disease detection and treatment may stop the spread of illnesses, minimize financial losses, and guarantee the healthy growth of the cotton sector. To address the issue of the absence of lightweight models for correctly recognizing cotton illnesses in the environment, we propose the CSGNet model for resource-constrained mobile devices, which is based on a tiny CNN architecture (ShuffleNet V2). Shuffle Attention (SA) is added to the model to improve disease feature extraction under complex background conditions. Additionally, GAUSSIAN ERROR LINEAR UNITS (GELU) are chosen to help the propose model better mine pertinent features, and the Adam optimization algorithm is used to enhance the model's generalization capabilities. The public data set of cotton leaf disease photos with a natural backdrop was utilized for experimental simulation training and performance testing. The identification accuracy was 99.1%, which was 1.5 percentage points better than previously, and the model size was just 4.96 MB. In comparison to other classification network models like MobileNet-V3, ResNet-50, and DenseNet-121, it not only achieves a greater recognition accuracy but also has a quicker convergence process and fewer parameters. CSGNet can achieve intelligent detection of cotton leaf diseases based on mobile end, and it complies with the practical application requirements. Even mobile devices with little computing power may use it effectively.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".